As computer systems have advanced, processing power and capabilities have increased both terms of general processing and more specialized processing such as graphics processing and chipsets. As a result, increasingly smaller and more powerful chips are constantly manufactured. As smaller more advanced chips are made, the number of transistors increases resulting in increased power consumption and heat production. This has made chip cooling important but has also made the packaging materials important as repeated heating and cooling during typical usage induces stresses on packages thereby effecting the life of the chip.
Before a chip package is finalized, the design is tested to ensure it is able to withstand stresses induced by repeated heating and cooling during usage. During semiconductor qualification, temperature cycling is done to simulate the temperature going up and down and thereby inducing thermo-mechanical failure. As the temperature goes up and down, there is a thermal mismatch between different materials in a package. For example, the thermal expansion coefficient of a silicon die and the thermal expansion coefficient of its substrate may be different. Thus, at higher temperatures the substrate may expand more than the die unfortunately causing stress at the interface between the die and the substrate.
Currently, accelerated temperature cycling (ATC) is used widely as a semiconductor test method to address thermo mechanical failure mechanisms likely to occur in actual field operations. ATC creates thermo-mechanical stress by alternating the chamber ambient temperature to reach designated high and low temperature extremes. Due to this ambient driven temperature control, the whole body of the testing package or system follows the ambient temperature. That is, the heating and cooling are done by heating and cooling the test chamber surrounding the package. One of the problems with this kind of traditional temperature cycling is that it is isothermal because the temperature is driven by the ambient temperature and the chip remains unpowered. Thus, each component within the package changes temperature at the same time.
Unfortunately, thermo-mechanical stresses from actual operating conditions are quite different from the ambient controlled isothermal nature of ATC. Under typical user conditions of power on/standby/off cycling of devices, the powered silicon (Si) die thereof acts as a heat source while other components act as heat dissipaters. This creates non-uniform heating and non-uniform temperature distribution in the package. This may make the device or system deform different than would be observed with ATC. Thus, when ATC testing is used the results do not sufficiently reflect real world conditions. That is, ATC is not able to simulate an actual user's power on/standby/off cycling conditions due to the nature of ATC being an isothermal ambient temperature controlled test method.
Two conventional solutions have attempted to mimic the non-isothermal temperature gradient of actual user conditions. One conventional method uses a thermal test chip which creates self heating when powered on. However, this method does not provide consistent and accurate temperature profiles across testing devices as thermal behavior of each device can be different. It may require a dedicated active heat sink for fast cooling to increase the number of cycles within a given time period. The testing temperature range is also limited as this method relies on self heating and ambient temperature cooling.
Another conventional method uses a real application system with a mounted testing device. For example, a testing device is soldered down to a real computer system. By turning the computer on and off (or power cycling), the testing device is self heated and ambiently cooled. However, this method also does not provide consistent and accurate temperature profiles across testing devices as thermal behavior of each device can be different. The testing temperature range and number of cycles within a given time period are also limited as this method relies on self heating and ambient temperature cooling. Further, this method involves high costs and time in building systems, developing special software, and maintenance.